Enhanced Secretory Production of a Single-Chain Antibody Fragment from Bacillus subtilis by Coproduction of Molecular Chaperones

Abstract

Formation of inclusion bodies is a major limiting factor for secretory production of an antidigoxin single-chain antibody (SCA) fragment from Bacillus subtilis. To address this problem, three new strains with enhanced production of molecular chaperones were constructed. WB600BHM constitutively produces the major intracellular molecular chaperones in an appropriate ratio without any heat shock treatment. This strain reduced the formation of insoluble SCA by 45% and increased the secretory production yield by 60%. The second strain, WB600B[pEPP], overproduces an extracytoplasmic molecular chaperone, PrsA. An increase in the total yield of SCA was observed. The third strain, WB600BHM[pEPP], coproduces both intracellular and extracytoplasmic molecular chaperones. This led to a further reduction in inclusion body formation and a 2.5-fold increase in the secretory production yield. SCA fragments secreted by this strain were biologically active and showed affinity to digoxin comparable to the affinity of those secreted by strains without overproduction of molecular chaperones. Interestingly, accumulation of a pool of periplasmic SCA was observed in the PrsA-overproducing strains. This pool is suggested to represent the secreted folding intermediates in the process of achieving their final configuration.

We have previously reported the successful secretory production of a biologically active antidigoxin single-chain antibody (SCA) fragment by using a six-extracellular-protease-deficient Bacillus subtilis strain (36). The secreted SCA is comparable in specificity and affinity to the parental 26-10 monoclonal antibody. The production yield is around 5 mg/liter in a shake flask culture. To improve the yield, factors limiting production should be identified. In this study, analysis of the distribution of this protein in B. subtilis shows that the secreted fraction represents only 23% of the total SCA fragments produced by the cell. The majority of the SCA protein (60%) is insoluble inside the cell. To produce soluble foreign proteins, various approaches have been developed including lowering the cultivation temperature (3), applying osmotic stress (4), changing hydrophobic residues to hydrophilic ones through protein engineering (2, 16), and fusing proteins to solubilizing partners (20). While some of these approaches work well in certain cases, the others require sequence alteration in proteins or the construction of fusions. It would be ideal to develop an engineered host strain that is better able to mediate the proper folding of proteins so that different foreign proteins can be produced in their authentic forms without sequence alteration. One promising approach is the coproduction of molecular chaperones that mediate protein folding, assembly, and secretion (8, 22, 33). As in Escherichia coli, B. subtilis has two series of general molecular chaperones (GroE and DnaK series). Structural genes for these molecular chaperones are organized into two operons, the groE operon (groES-groEL), and the dnaK operon (hrcA-grpE-dnaK-dnaJ-orf35-orf28-orf50) (13, 23). Studies from E. coli indicate that these two series of molecular chaperones can act either independently or synergistically in a successive manner to facilitate the proper folding and assembly of certain proteins (9, 12, 19). Overproduction of only one series of molecular chaperones showed variable results in increasing protein solubility (22, 33). This observation reflects the difficulty in predicting which set of molecular chaperones is the limiting factor for foreign protein production. Moreover, overproduction of individual molecular chaperones without the cochaperones can even be toxic to the host (5). Hence, it is desirable to have a balanced coproduction of both series of molecular chaperones in the same expression host. With the characterization of the regulatory regions from the B. subtilis groE (41) and dnaK operons (42), these two operons were found to be regulated by a common repressor, HrcA, with its activity modulated by GroE (26). Inactivation of hrcA results in the constitutive production of intracellular molecular chaperones from these two operons (40). In this study, a new six-extracellular-protease-deficient B. subtilis strain (WB600BHM) with hrcA inactivated was constructed. Furthermore, we also modulate the expression level of an extracytoplasmic molecular chaperone, PrsA, in B. subtilis (15). prsA from B. subtilis has been cloned and characterized by Kontinen et al. (17). It encodes a 33-kDa lipoprotein bound to the outer surface of the cell membrane. This lipoprotein is suggested to mediate protein folding at the late stage of secretion. Overproduction of PrsA has been shown to increase the production of α-amylase and proteases (18). With the WB600BHM strain and a binary plasmid vector system that overproduces PrsA and SCA, effects of coproduction of these molecular chaperones on the production of the antidigoxin SCA fragment were studied. Roles of bacterial cell wall and PrsA on the release of the secreted SCA fragments were also examined.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

B. subtilis WB600 (35) is a six-extracellular-protease-deficient strain with the nprB gene (34) disrupted by an erythromycin resistance marker. WB600B (trpC2 nprE ΔaprE Δbpf Δepr mpr::ble nprB::bsr) (39) is a derivative of WB600. Its nprB gene was disrupted by a blasticidin S resistance marker (bsr). It was used as the expression host in this study and also as the parental strain for the construction of WB600BHM. Plasmid pATD2 (36) is a pUB110 derivative carrying the structural gene for the antidigoxin SCA under control by the P43 promoter (36). Secretion of this protein is directed by the signal sequence from B. subtilis levansucrase. Plasmid pUB18-P43 was used here as a negative control. It is identical in sequence as pATD2 except for the absence of the structural gene encoding the signal sequence and SCA fragment. Cells carrying plasmids were cultivated in superrich medium (36) supplemented with the appropriate antibiotics for 5 to 6 h at 37°C. Final concentrations of antibiotics were as follows: kanamycin, 10 μg/ml; gentamicin, 2 μg/ml; and erythromycin and lincomycin, 5 μg/ml for each.

Construction of pE18 and pEPP.

Plasmid pE18 is a derivative of pE194-cop6 (14). It was constructed stepwise as follows. pE194-cop6 was first digested by AccI, followed by the fill-in reaction and the addition of a HindIII linker. By the same approach, the NdeI site in the resulting plasmid was also changed to an EcoRI site to generate pEH. A HindIII-EcoRI polylinker sequence from pUC18 was inserted in pEH to generate pE18. Plasmid pEPP is a pE18 derivative carrying a P43-prsA cassette that overproduces PrsA constitutively. A 300-bp EcoRI-KpnI fragment carrying the P43 promoter (36) was inserted in pE18 to generate pE18-P43. Structural gene for PrsA was amplified from B. subtilis 168 by PCR using Vent polymerase with the forward (5′ GGGGTACCGAATGATTAGGAGTGTTTG 3′) and backward (5′ GGGGATCCAAAAAAAGCTGTGCGCTCAATG 3′) primers. A KpnI site was introduced at the 5′ end and a BamHI site at the 3′ end of the amplified fragment. The resulting 960-bp fragment was then digested by KpnI and BamHI and inserted into pE18-P43 to generate pEPP.

Protein production studies.

B. subtilis strains carrying expression plasmid(s) were cultivated in superrich medium and harvested by centrifugation. The culture supernatant was concentrated by trichloroacetic acid (TCA) precipitation to yield the secreted protein fraction. To confirm the quantitative recovery of SCA fragments in the medium by TCA precipitation, affinity-purified SCA fragments with known quantity were either loaded directly onto a sodium dodecyl sulfate (SDS)-polyacrylamide gel or added to the secreted fraction of the negative control and collected via TCA precipitation for electrophoresis. Recovery of SCA fragments via TCA precipitation was confirmed to be in the range of 95% by Western blot analysis. To prepare intracellular soluble and insoluble protein fractions, cell pellets were washed once in phosphate-buffered saline (PBS), resuspended in cell disruption buffer (30 mM Tris HCl, 100 mM NaCl, 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride [pH 7.4]) and lysed in a French press. The insoluble fraction was collected by centrifugation at 20,000 × g. Protein samples were normalized against the cell density and analyzed on an SDS–10% polyacrylamide gel. For the Coomassie blue-stained gel shown in Fig. 1, proteins corresponding to 400 μl of culture for the secreted fraction, 40 μl for the soluble fraction, and 200 μl for the insoluble fraction were loaded. For Western blot analyses, proteins corresponding to 400 μl of culture for the secreted fraction and 200 μl each for soluble and insoluble fractions were loaded. The amount and ratio of proteins loaded were selected to ensure that SCA fragments from different fractions could be effectively detected. The amount was also chosen to ensure that the proteins could be completely transferred to the nitrocellulose membrane and that the intensity of the blotted SCA bands probed by rabbit antiserum (36) specifically against the antidigoxin SCA fragment was in the linear range for quantification. Images were taken with the GDS 7500 gel documentation system from UVP. Quantification of SCA produced from the digitized images was done on the Western blot (Fig. 1B), with a Fuji bioimaging analyzer system (BAS 1000, Fuji Photo Film Co., Ltd.) and MacBAS software.

FIG. 1.

Analysis of effects of molecular chaperones on antidigoxin SCA distribution in various production strains. (A) Coomassie blue-stained gel; (B) Western blot using SCA-specific antiserum. Lanes: 1 to 4, secreted fraction; 5 to 8, soluble fraction (S); 9 to 12, insoluble fraction (I); 1, 5, and 9, WB600B[pATD2]; 2, 6, and 10, WB600BHM[pATD2]; 3, 7, and 11, WB600B[pATD2, pEPP]; 4, 8, and 12, WB600BHM[pATD2, pEPP]. a, DnaK (lanes 6 and 8); b, GroEL; c, GroES; arrow, secreted 31-kDa SCA; arrowhead, SCA precursor; open circle, PrsA. Amounts of proteins loaded were as described in Materials and Methods.

Protoplast preparation.

Protoplasts were prepared as described by Merchante et al. (25), with minor modifications. Cells (10 ml) were harvested by centrifugation at 7,000 × g at 20°C. The culture supernatant was collected as the secreted fraction. The cell pellets were washed twice with SET buffer (20% sucrose, 50 mM Tris HCl, 50 mM EDTA [pH 7.6]) and resuspended in 2 ml of protoplast buffer (66% sucrose, 50 mM Tris HCl, 16 mM MgCl₂ [pH 8.0]). The suspension was incubated at 37°C for 45 min in the presence of lysozyme (400 μg) and phenylmethylsulfonyl fluoride (1 mM, final concentration). The mixture was then centrifuged at 21,000 × g (15 min, 20°C) to yield the periplasmic fraction (supernatant) and protoplast pellet. Protoplasts were resuspended in lysis buffer (50 mM Tris HCl, 5 mM MgSO₄ [pH 8.0]) and disrupted in a French press. The insoluble fraction was collected by centrifugation at 20,000 × g. As a control, the cell pellet from a similarly prepared 10-ml cell culture was directly resuspended in lysis buffer without lyzozyme treatment and disrupted in a French press to yield the cellular soluble and insoluble fractions.

Glucose 6-phosphate dehydrogenase assay.

Reduction of NADP to NADPH was monitored spectrophotometrically at A₃₄₀ as described by Merchante et al. (25). An appropriate volume of each fraction (equivalent to 0.02 to 0.3 ml of cell culture) in a final volume of 1 ml of reaction mixture was used in the assay. The reaction was initiated by the addition of glucose 6-phosphate, and the change of A₃₄₀ was monitored continuously at room temperature for 15 min. The initial reaction rate in the linear range was used for the calculation. The background dehydrogenase activity, measured on a blank mix which contained all components except glucose 6-phosphate, was subtracted for each cell fraction.

Affinity purification of antidigoxin SCA.

An ouabain-Sepharose column was prepared as described previously (36). Cells were pelleted by centrifugation at 7,000 × g for 10 min. Culture supernatant or periplasmic fraction was applied to a 1.5-ml ouabain-Sepharose column equilibrated with PBS. The column was washed extensively with PBS, and bound material was eluted with 20 mM ouabain. The eluant was concentrated by centrifugation through a Centricon-10 (Amicon). The purity of the eluted antidigoxin SCA was analyzed by electrophoresis on an SDS–10% polyacrylamide gel.

Estimation of the functional yield of antidigoxin-SCA.

One milliliter of culture supernatant was applied to a 1.5 ml ouabain-Sepharose column. The affinity-purified SCA was recovered in a total volume of 1 ml of PBS. Then 20 μl of this eluant and 20 μl of culture supernatant were analyzed by Western blotting.

N-terminal sequence analysis of periplasmic antidigoxin SCA.

The affinity-purified antidigoxin SCA fragment from the periplasmic fraction was spotted directly to an Immobilon membrane for N-terminal sequencing (36). The sequence for the first five amino acids from the periplasmic SCA was determined with an Applied Biosystems 473 pulsed liquid protein sequencer at the Protein Microchemistry Centre, University of Victoria.

Kinetic parameters of the secreted SCA fragments determined by surface plasmon resonance.

Kinetics of digoxin-antidigoxin SCA interaction was studied in real time, using a BIAcoreX biosensor from Biacore, Inc. (Piscataway, N.J.). Digoxin was first coupled to bovine serum albumin (BSA) as described by Smith et al. (32). The complex was then covalently immobilized to carboxylmethyl groups of dextran on the surface of the CM5 sensor chip via the amine coupling method specified by the manufacturer. The immobilization level was around 600 resonance units. The reference cell consists of BSA immobilized on a CM5 sensor chip. To determine the association rate constant (on rate; k_a), culture supernatants (containing secreted antidigoxin SCA) at various dilutions in HBS buffer (10 mM HEPES [pH 7.4], 150 mM NaCl, 3.4 mM EDTA, 0.05% BIAcore surfactant P20) were injected over the sensor chip surface at a flow rate of 40 μl/min. Contact time between the antigen and antibody was maintained at 2.25 min. According to the kinetic equation dR/dt = k_aCR_max − (k_aC + k_d)R (27), a plot of the binding rate (dR/dt) versus response (R; expressed in terms of resonance units) will be linear to give a slope, k_s, which is defined as k_s = k_aC + k_d. The plot of k_s, generated with different concentrations (C) of antidigoxin-SCA, versus C will yield a straight line with a slope that determines the value of k_a. The antidigoxin SCA concentration in the culture supernatant was estimated from the standard curve established with purified, electroblotted antidigoxin SCA probed against specific antibody. Dissociation of antibody from the surface was monitored for about 10 min with continuous flow of HBS buffer, and the dissociation rate constant (off rate; k_d) was determined according to the kinetic equation (27), ln (R₀/R_t) = k_d (t − t₀). A plot of ln (R₀/R_t) versus (t − t₀) again yields a straight line with the slope that reflects the value of k_d. Surface regeneration between cycles was achieved by a 2-min pulse of 15 mM HCl at 5 μl/min.

RESULTS

Development of a host-vector system for coproduction of both intracellular and extracytoplasmic molecular chaperones.

To construct a six-extracellular-protease-deficient B. subtilis strain that constitutively overproduces both the GroE and DnaK series of molecular chaperones without heat shock, the repressor gene, hrcA, which controls the expression of these operons was inactivated. pKS-3 is an integration plasmid that carries a truncated hrcA with an inserted gentamicin resistance marker derived from pGE1 (6). By introducing the KpnI-linearized pKS-3 to WB600B and selecting for gentamicin resistance, one of the resulting strains, designated WB600BHM (HM stands for hyperproduction of molecular chaperones), was further characterized. PCR amplification of hrcA from both WB600B and WB600BHM confirmed the inactivation of hrcA in WB600BHM (data not shown). The growth of WB600BHM in superrich medium is similar to that of WB600B. Western blot analysis showed that the GroEL and DnaK levels in WB600BHM are 7- and 2.5-fold, respectively, higher than those in WB600B, (data not shown). With the construction of an expression vector (pEPP) to overproduce PrsA, the effects of intracellular molecular chaperones and PrsA, used singly or together, on the production and secretion of SCA were studied systematically. Distribution of SCA in the secreted and intracellular fractions of different constructs was determined by SDS-polyacrylamide gel electrophoresis and Western blotting (Fig. 1 and Table 1).

TABLE 1.

Effects of coproduction of molecular chaperones on antidigoxin SCA production and distribution^a

Strain	Relative amt of total SCA produced	Distribution (%)
		Secreted fraction	Cellular fraction
			Soluble	Insoluble
WB600B[pATD2]	1.0	23 (1.0)b	17 (1.0)	60 (1.0)
WB600BHM[pATD2]	1.0	37 (1.6)	30 (1.7)	33 (0.55)
WB600B[pEPP, pATD2]	1.7	22 (1.6)	38 (3.7)	40 (1.1)
WB600BHM[pEPP, pATD2]	1.3	43 (2.5)	51 (3.9)	6 (0.13)

^aQuantification was based on the Western blot analysis in Fig. 1B. The amount of SCA in each fraction has been corrected for difference in the amount of samples loaded. 

^bAmount of SCA in fraction relative to the amount in strain WB600B[pATD2]. 

Effects of coproduction of intracellular molecular chaperones on SCA production.

With the absence of coproduction of any molecular chaperones in WB600B[pATD2], the majority of SCA was found to be insoluble in the intracellular fraction (Fig. 1, lane 9). Coproduction of intracellular molecular chaperones (DnaK, GroEL, and GroES [Fig. 1A, lane 6]) in WB600BHM[pATD2] changed this pattern. Although the total production yield of SCA in this strain was the same as in WB600B[pATD2], the distribution of SCA in different fractions was shifted. SCA in the intracellular insoluble fraction decreased from the original 60% to 33% (Table 1). On the other hand, secreted SCA increased by 60% and SCA in the intracellular soluble fraction by 70%. The majority of SCA in the intracellular soluble fraction existed in the precursor form (Fig. 1B, lane 6).

Effects of coproduction of PrsA on SCA production.

The binary plasmid vector system (pEPP and pATD2) allows the coproduction of PrsA and SCA in the same host. By using WB600B as the host, this system constitutively overproduces PrsA, a 33-kDa protein detected predominantly in the insoluble fraction (Fig. 1A, lane 11). The amount of PrsA produced at 4 to 6 h of culture was estimated to be over eightfold that of WB600B[pATD2]. In WB600B[pEPP, pATD2], the total amount of SCA produced increased by 70%. This increase is mainly due to the increase of SCA in both the secreted fraction (1.6-fold) and the intracellular soluble fraction (3.7-fold). SCA in the soluble fraction was mainly in the mature form (Fig. 1B, lane 7).

Effects of coproduction of both intracellular and extracytoplasmic molecular chaperones.

WB600BHM[pEPP, pATD2] overproduces both PrsA and the major intracellular molecular chaperones. Production of GroESL and DnaK was fairly constant throughout the cell cycle. Although the production of PrsA increased continuously well past the early stationary phase when the culture was harvested, the production level of SCA reached its maximum at this stage. The presence of these molecular chaperones apparently had minimal negative effect, if any, on cell growth. In comparison with WB600B[pATD2], the total amount of SCA produced by WB600BHM[pEPP, pATD2] increased by 30%. The secreted SCA was up 2.5-fold, and the intracellular soluble fraction nearly quadrupled. But most importantly, the amount of insoluble SCA in the intracellular fraction decreased significantly (Fig. 1, lane 12; Table 1). In the pool of soluble SCA in the intracellular fraction, the levels of both the precursor and mature forms of SCA increased with the mature form predominant (Fig. 1B, lane 8).

Quantification of secreted SCA fragments.

To confirm that the coproduction of both intracellular and extracytoplasmic molecular chaperones can increase the secretory production yield of SCA by at least 2.5-fold, SCA secreted from WB600B[pATD2] was affinity purified to homogeneity and used as the standard for quantitative Western blot analysis. As shown in Fig. 2 (lanes 1 to 3), the intensity of the blotted SCA bands showed a linear response in the range from 25 to 200 ng. The average production yields of secreted SCA from WB600B[pATD2] and WB600BHM[pEPP, pATD2] from three independent experiments were determined to be 4 and 12 mg/liter, respectively. Therefore, coproduction of these molecular chaperones results in a threefold increase in the secretory yield of SCA.

FIG. 2.

Western blot analysis for quantification of SCA fragments and measurement of the antigen binding ability of SCA fragments. Lanes 1 to 3 represent the loading of 25, 100, and 200 ng of the affinity-purified antidigoxin-SCA fragments, respectively. C (lanes 4 and 5) and IE (lanes 6 and 7) represent the secreted SCA fragments from WB600B[pATD2] and WB600BHM[pEPP, pATD2], respectively. B and A indicate the analysis of the samples (20 μl) before and after, respectively, loading to the ouabain-Sepharose column.

Biological activity of the secreted SCA fragments.

A recent study on the biosynthesis of an antihemagglutinin SCA fragment in a cell-free translation system indicates that the increase in solubility associated with the addition of molecular chaperones did not result in a corresponding increase in the antigen binding activity (30). It is hence vital for us to determine whether antidigoxin SCA fragments produced from WB600BHM[pEPP, pATD2] are all biologically active. Quantification of recovery of affinity-purified SCA fragments by using a Fuji bioimaging analysis system in Fig. 2 (lane 5 versus lane 4 for SCA from WB600B[pATD2]; lane 7 versus lane 6 for SCA from WB600BHM[pEPP, pATD2]) shows that the levels of recovery of active SCA fragments from WB600B[pATD2] and WB600BHM[pEPP, pATD2] were 85 and 93%, respectively. With multiple runs of SCA from WB600BHM[pEPP, pATD2] to the affinity matrix, the recovery was consistently in the range of 90 to 100%. Since a low percentage of loss of SCA over the affinity matrix is expected, these data suggest that essentially all the SCA fragments secreted by WB600BHM[pEPP, pATD2] are biologically active.

Kinetic parameters of interactions between antidigoxin SCA and digoxin.

Although antidigoxin SCA fragments produced by WB600BHM[pEPP, pATD2] are biologically active, it is important to determine whether they have the same affinity to digoxin as the SCA fragments produced by WB600B[pATD2]. Traditional methods (1, 36) allow the determination of the affinity constant without providing any information about the on rate and off rate of the interaction. By using the surface plasmon resonance technology (27), interaction between the antidigoxin SCA fragments and digoxin could be monitored in real time and both on and off rates could be determined. Culture supernatants containing SCA at different dilutions were injected to the flow cell which contains digoxin-BSA conjugate immobilized on a CM5 sensor chip. Binding of SCA fragments to digoxin was monitored as the resonance units changed with time. As shown in Fig. 3A, the plot of k_s versus different concentrations of SCA fragments yielded a straight line with a slope representing the on rate, k_a. SCA fragments from WB600B[pATD2] and WB600BHM[pEPP, pATD2] had similar slopes in this study, reflecting that they have the same on rate. To monitor dissociation, buffer was injected and the dissociation of the bound SCA-digoxin complex was reflected by the decrease in resonance units with time. The slope of the plot of ln (R₀/R_t) versus (t − t₀) reflects the off rate, k_d. These SCA samples again showed identical off rates. Therefore, SCA fragments produced from WB600BHM[pEPP, pATD2] have the same kinetic parameters as SCA fragments produced from WB600B[pATD2]. The apparent affinity constant (K_a, equivalent to k_a/k_d) of SCA fragments to digoxin as determined by the biosensor method is around 1.59 × 10⁹ M⁻¹. This value is comparable to that determined by other methods (36).

FIG. 3.

Interaction between antidigoxin SCA fragments and digoxin as monitored via the BIAcore biosensor. (A) Determination of the association rate constant (on-rate; k_a); (B) determination of the dissociation rate constant (off-rate; k_d). R₀ is the response at an arbitrary starting time t₀; R_t is the response at time t. Open and closed squares represent samples containing secreted antidigoxin SCA fragments from WB600B[pATD2] and WB600BHM[pEPP, pATD2], respectively.

Presence of mature SCA fragments in the periplasmic fraction in strains overproducing PrsA.

B. subtilis overproducing PrsA significantly enriched the SCA fragments in the soluble intracellular fraction (Fig. 1B, lanes 7 and 8; Fig. 4B, lane 2). At least 60% of these SCA fragments are equivalent in size to the mature form of SCA in the secreted fraction. We speculate that this form of SCA actually represents the mature SCA fragments that have translocated across the membrane but have not been released to the culture medium. To test this idea, protoplasts were prepared from WB600BHM[pATD2] and WB600BHM[pEPP, pATD2]. As shown in Fig. 4B (lane 4), this form of SCA fragments from WB600BHM[pEPP, pATD2] could be released to the medium once the cell walls were removed by lysozyme treatment. Most of the soluble SCA remaining in the protoplasts had the size equivalent to that of the precursor (lane 5). To eliminate the possibility that the release of mature SCA fragments to the periplasmic space is caused by cell lysis, distribution of a cytoplasmic marker, glucose 6-phosphate dehydrogenase, in different fractions was determined. No glucose 6-phosphate dehydrogenase activity could be detected in the secreted fraction. Around 95% of activity rested with the soluble fraction with the remaining 5% in the periplasmic fraction. Similar results were obtained whether WB600B[pATD2] or WB600BHM[pEPP, pATD2] was used. Thus, the degree of cell lysis generated during protoplast preparation was negligible. These data support the idea that over 60% of SCA fragments previously associated with the intracellular soluble fraction are actually periplasmic in nature and that these SCA fragments represent the form that has been secreted. To determine whether these mature SCA fragments have any antigen binding activity, the periplasmic fraction was passed through the ouabain affinity matrix. We found that 45% of the SCA fragments were retained on the column and selectively eluted off by ouabain, while the rest were in the flowthrough. This result indicates that 55% of the periplasmic SCA fragments have not achieved their active configuration. To confirm that the SCA fragments in the periplasmic fraction have the signal sequence processed, the sequence of the first five amino acids of the affinity-purified SCA fragments in this fraction was determined. The sequence was found to be DVVMT, which matches the N-terminal sequence of the SCA light chain. Although the N-terminal sequence of the inactive SCA fragments in the periplasmic fraction was not determined, their signal sequences were expected to be processed as well, since active and inactive SCA fragments comigrated as a single band in the SDS-polyacrylamide gel.

FIG. 4.

Distribution of antidigoxin SCA fragments in B. subtilis WB600BHM[pATD2] and WB600BHM[pEPP, pATD2] by Western blot analysis. IF, intracellular fractions; Sec, secreted fraction (extracellular); Sol and Ins, soluble and insoluble intracellular fractions, respectively; Per, periplasmic fraction. + or − lysozyme indicates whether the samples have been treated with lysozyme or not. Amounts of protein loaded are normalized to cell density across all lanes.

DISCUSSION

Formation of inclusion bodies is a common problem in producing foreign proteins in a heterologous expression system. In our initial study, most of the antidigoxin SCA fragments produced in B. subtilis were found to be in the insoluble fraction. We have tried to increase solubility of the SCA by overproducing thioredoxin. However, unlike what has been reported for E. coli (37), this approach did not work in our case. To address the problem, a B. subtilis strain, WB600BHM, which overproduces the major intracellular molecular chaperones, and a vector system which overproduces an extracytoplasmic molecular chaperone were constructed. WB600BHM offers several advantages as a production host. (i) Both sets of major intracellular molecular chaperones (GroE and DnaK series) are overproduced at a moderate level; extra energy can thus be reserved for the overproduction of foreign proteins. (ii) Individual molecular chaperones (i.e., GroES with GroEL and DnaK with DnaJ/GrpE) are produced in an appropriate ratio. Cell toxicity so frequently associated with an unbalanced overproduction of a molecular chaperone (5, 23) was not observed in our study. (iii) The chromosomal genes for these intracellular molecular chaperones are stable; no extra plasmid is required for their expression. The second plasmid in a binary plasmid system can thus be applied to express other accessory factors (e.g., PrsA in this case). (iv) The two heat-inducible lon protease genes in B. subtilis (29, 38) are not regulated by HrcA. Therefore, these proteases should not be induced because of the inactivation of hrcA. In contrast, heat-inducible proteases such as Lon and ClpP are expected to be produced in E. coli when ς³² is produced from an artificial system to increase the expression levels of intracellular molecular chaperones without heat shock (33).

The major effects of coproduction of intracellular molecular chaperones on SCA production are the minimization of SCA aggregation and the increase of the secretory production yield of SCA. These observations are consistent with some of the findings of the effects of the same molecular chaperones on foreign protein production in E. coli (22, 33).

The major effect of the extracytoplasmic molecular chaperone, PrsA, on SCA production was an increase in the total amount of SCA produced in all fractions, particularly in the soluble intracellular one. Our fractionation and protein sequencing studies demonstrated that there were two forms of SCA fragments in the soluble intracellular fraction: the precursor form and the mature form. While the precursor form was located solely inside the cell, the mature fragments had actually been translocated across the membrane since their signal sequence had been cleaved, and they could be released to the medium once the cell walls were removed. With the recent demonstration of an operationally defined periplasm in B. subtilis (25), we suggest that these processed SCA fragments represent the folding intermediates transiently trapped between the cell wall and cell membrane. This idea is strengthened by the observation that at least 55% of these SCA fragments were biologically inactive. Their presence in the periplasm could result from several possibilities. One possibility could arise from their interaction with PrsA. For this, we propose that B. subtilis PrsA functions as an extracytoplasmic molecular chaperone in a manner which is similar to what SurA does in E. coli (21). Both PrsA and SurA show good homology to parvulin, a new class of peptidyl-prolyl cis-trans isomerase from E. coli (28). This enzyme catalyzes isomerization of prolyl residues, which can be a rate-limiting step in protein folding. Under in vitro conditions, peptidyl-prolyl cis-trans isomerase has been shown to catalyze proline isomerization of Mab33 (11) and slightly improve the production yield of a SCA-toxin fusion (7). Based on the three-dimensional structure of the 26-10 Fab fragment, the antidigoxin SCA is expected to have two cis-proline residues. Therefore, PrsA can potentially interact with these folding SCA intermediates before they gain the final active configuration. Alternatively, the processed form may represent loosely folded structures which have difficulty in penetrating the bacterial cell walls.

Coproduction of PrsA also resulted in an increase in the secretory production yield of SCA. This can be explained by the greater stability of the secreted SCA in the presence of PrsA. Although we used a six-extracellular-protease-deficient strain as the production host, the presence of another extracellular protease (Vpr [31]), a wall-bound protease (WprA [24]), and even some putative extracellular proteases identified through genomic sequencing (10) can affect the stability of the secreted SCA fragments. Proper folding of SCA may allow the protein to gain resistance against proteases and hence increased stability in the secreted fraction. This view is supported by the findings that exoenzymes in a prsA mutant (15) and several outer membrane proteins in surA mutants (21) are degraded rapidly.

With the combination of both intracellular and extracytoplasmic molecular chaperones, the problem of inclusion body formation for this antidigoxin SCA fragment has been reduced substantially. If we include the mature form of periplasmic SCA into the secreted fraction, over 75% of the total SCA fragments synthesized has actually been exported outside the cells. Although not all the secreted SCA fragments are released into the medium, this system operates in such a way that only high-quality, biologically active SCA fragments can be harvested from the culture supernatant.